Physician Judgment and Machine Predictions

The study goal is to improve the value of care and reduce health disparities by developing a targeted set of sophisticated and powerful algorithms to improve upon human clinical judgments. The plan is to use the test case of detecting sepsis in patients in the emergency department (ED) as the first step in improving the value of care and reducing health disparities by developing a targeted set of sophisticated and powerful algorithms to improve upon human clinical judgments. This work will be performed using data from the University of Pennsylvania Health System where a preliminary Early Warning and Response System for Sepsis monitors clinical parameters. The premise underlying all this work is that by improving decision-making, it will both reduce low-value care and health disparities.